This was the third week was of test soundings in the northwest corner of the sanctuary dedicated to Hadrian and Antoninus Pius, and we made further discoveries about the sanctuary's occupation history:

Phase 1: Construction of the sanctuary during the mid- to second half of the second century A.D. was further corroborated by a fill on the weathered bedrock, which once had supported the courtyard's pavement. We found material of the same date inside the portico, where two walls of a possible cellar below the north portico were located some two meters below the level of the portico's steps. This presumed cellar was made of mortared rubble built against the lower part of the foundations of the northern temenos (sanctuary) wall.

Phase 2: Of the two phases of encroachment on the temple complex, the first may have taken place during the fifth century. Inside the portico we found a late Roman floor level (late-fifth to early-sixth century) well below the level of the portico's lowest step. It belonged to a room between the steps and the portico's back. This floor level was on top of the second-century fill. The walls of this building phase incorporate many elements of the portico proper, which therefore must already have been dismantled or destroyed (by the A.D. 500 earthquake?).

The second encroachment could be dated to the early Byzantine period (sixth century). It consisted of walls built on structures of the first encroachment phase. These included a 1.04 m wide door, which seems to confirm that one is dealing here with a series of rooms along the same line as that of the original portico's steps. These structures, however, also extend beyond the line of the portico's original back wall.

Phase 3: During the middle Byzantine period (eleventh century, again confirmed by an anonymous follis dated to the years 1078-1081), material was dumped here, inside the remains of the second encroachment phase, together with earlier refuse. Moreover we established that there were some walls atop this dumped material, with an orientation similar to that of the earlier structures. Our hypothesis of a medieval kastron for this last occupation phase still stands (see report July 20-26).